Transcript
Oversampling Converters and PLLsTuesday 23rd of March, 2010, 9:15 – 11:10uesday 3 d o a c , 0 0, 9 5 0
Snorre Aunet, sa@ifi.uio.noNanoelectronics Group, Dept. of Informatics
Office 3432
Last time – and today, Tuesday 16th of March:
L iLast time:
13.6 Interpolating A/D Converters p g13.7 Folding A/D Converters 14.1 Oversampled converters
Today:
14.2 Oversampling with noise shaping14 3 S t A hit t14.3 System Architectures14.4 Digital Decimation Filters14.5 Higher-Order Modulators(14 6 Bandpass O ersampling Con erters)(14.6 Bandpass Oversampling Converters)14.7 Practical Considerations14.8 Multi-bit oversampling converters
2nd order sigma delta design example2nd order sigma delta design example 16.1 Basic Phase Locked Loop Architecture
Nyquist Sampling, Oversampling, Noise Shaping Fi fShaping • Figure from
[Kest05]• Straight over-Straight over
sampling gives an SNR i timprovement of 3 dB / octave
• fs > 2f0 (2f0 = Nyquist Rate
• OSR = fs/2f0• SNRmax =
6 02N+1 76+6.02N+1.76+10log (OSR)
OSR, modulator order and Dynamic Range2 X increase• 2 X increase in M (6L+3)dB or(6L+3)dB or (L+0.5) bit increase inincrease in DR.L i d lt• L: sigma-delta orderO li• Oversampling and noise h ishaping
14.2 Oversampling with noise shaping
• The anti aliasing filter bandlimits the input signals less than fs/2.g p g s• The continous time signal xc(t) is sampled by a S/H (not
necessary with separate S/H in Switched Capacitor impl.)• The Delta Sigma modulator converts the analog signal to a
noise shaped low resolution digital signalTh d i h l d l l i di i l• The decimator converts the oversampled low resolution digital signal into a high resolution digital signal at a lower sampling rate usually equal to twice the desired bandwidth of the desiredrate usually equal to twice the desired bandwidth of the desired input signal (conceptually a low-pass filter followed by a downsampler).p )
Noise shaped Delta Sigma Modulator
23. mars 2010 7
First-Order Noise Shaping (Figures from Schreier & Temes ’05)
• STF(z) = [H(z)/1+H(z)] (eq. 14.15) NTF(z) = [1/1+H(z)]• Y(z) = S (z) U(z) + N (z) E(z)• Y(z) = STF(z) U(z) + NTF(z) E(z)• H(z) = 1/z-1 (discrete time integrator) gives 1st order noise shaping • STF(z) = [H(z)/1+H(z)] = 1/(z-1)/[1+1/(z-1)] = z-1STF(z) [H(z)/1 H(z)] 1/(z 1)/[1 1/(z 1)] z• NTF(z) = [1/1+H(z)] = 1/[1+1/(z-1)] = ( 1 – z-1)• The signal transfer function is simply a delay, while the noise
transfer function is a discrete-time differentiator (i.e. a high-pass filter)
14.2 Oversampling with noise shaping
Quantization noise power for linearized model of a general ΔΣ modulator
23. mars 2010 10
Second-order noise shaping
14.3 System Architectures (A/D) • Xc(t) is sampled and held, resulting in x (t)resulting in xsh(t).
• xsh(t) is applied to an A/D Sigma Delta modulator which has a 1-bit output, xdsm(n). p dsm( )The 1-bit signal is assumed to be linearly related to the input Xc(t) (accurate to many orders of resolution)orders of resolution), although it includes a large amount of out-of-band quantization noise (seen to q (the right).
• A digital LP filter removes any high frequency content, i l di t f b dincluding out of band quantization noise, resulting in Xlp(n)
• Next Xl (n) is resampled atNext, Xlp(n) is resampled at 2f0 to obtain Xs(n) by keeping samples at a submultiple of the OSR
System Architectures (D/A) • The digital input, Xs(n) is a multi-bit signal
and has an equivalent sample rate of 2fand has an equivalent sample rate of 2f0,where f0 ,is slightly higher than the highest input signal frequency.
• Since Xs(n) is just a series of numbers the frequency spectrum has normalized thefrequency spectrum has normalized the sample rate to 2л.
• The signal is upsampled to an equivalent higher sampling rate, fs, resulting in the signal xs2(n)g s2( )
• xs2(n) has images left that are filtered out by the interpolation filter (brick-wall type) to create the multi-bit signal Xlp(n) , by digitally filtering out the images.g y g g
• Xlp(n) is applied to a fully digital sigma delta modulator producing the 1-bit signal, Xdsm(n) , containing shaped quantization noise.
• Xdsm(n) is fed to a 1-bit D/A producing Xda(t), which has excellent linearity properties but still quantization noise.
• The desired signal, Xc(t) can be obtained by using an analog filter to filter out the out-of-band quantization noise. (filter should be at least one order higher than the modulator.)
14.4 Digital decimation filters• Many techniques• Many techniques• a) FIR filter removes much of
the quantization noise, so that the output can bethe output can be downsampled by a 2nd stage filter which may be either IIR type (as in a) uppermost ) ortype (as in a), uppermost ) or a cascade of FIR filters (as in b), below )
• In b) a few halfband FIR filtersIn b) a few halfband FIR filters in combination with a sinc compensation FIR-filter are used. In some applications, these halfband and sinc compensation filters can be prealized using no general multi-bit multipliers [Saramaki, 1990]
14.5 Higher-Order Modulators –Interpolative structureInterpolative structure
Lth d i h i d l t i SNR b 6L 3dB/ t• Lth order noise shaping modulators improve SNR by 6L+3dB/octave.• Typically a single high-order structure with feedback from the quantized
signal. • In figure 14.20 a single-bit D/A is used for feedback, providing excellent
linearity.• Unfortunately modulators of order two or more can go unstable especially• Unfortunately, modulators of order two or more can go unstable, especially
when large input signals are present (and may not return to stability) Guaranteed stability for an interpolative modulator is nontrivial.
Multi-Stage Noise Shaping architecture (”MASH”)( MASH )
• Overall higher order modulators are constructed using lower order more• Overall higher order modulators are constructed using lower-order, more stable, ones more stable overall system.
• Fig. 14.21: 2nd order using two first-order modulators.• Higher order noise filtering can be achieved using lower-order modulators.• Unfortunately sensitive to finite opamp gain and mismatch
14.7 Practical considerations
•Stability•Linearity of two level converters•Linearity of two-level converters• Idle tones•Dithering•Opamp gain•Opamp gain
Design example, 14b 2nd order Sigma-Delta mod
• 16 bit, 24 kHz , OSR as powers of two, and allowing for increased baseband noise due to nonidealities: OSR 512 was chosen
Design example, 14b 2nd order Sigma-Delta mod
A t l t id liti• Among most relevant nonidealities:• Finite DC gain
Bandwidth• Bandwidth,• Slew rate• Swing limitation• Swing limitation• Offset voltage• Gain nonlinearityGain nonlinearity• Flicker noise• Sampling jitterSampling jitter• Voltage dependent capacitors• Switch on-resistance• Offset voltage and settling time for
comparators
Design example, 14b 2nd order Sigma-Delta mod
• Noninverting parasitic insensitive integrator (fig 10.9) was used (fully differential implementation)
2nd order modulator; top level schematics
• Two-phase clock generator, switches, chopper stabilized OTA (1st int.), OTA (2nd int.- fully differential folded cascode), comparator, latch, two-level DAC. Biasing circuit. Functional after test.
Phase-locked loops (chapter 16) • Phase-locked loopPhase locked loop• From Wikipedia, the free encyclopedia• A phase-locked loop or phase lock loop (PLL) is a control system that
t i l th t h fi d l ti t th h f " f "generates a signal that has a fixed relation to the phase of a "reference" signal. A phase-locked loop circuit responds to both the frequency and the phase of the input signals, automatically raising or lowering the frequency of a controlled oscillator until it is matched to the reference in both frequency and phase. A phase-locked loop is an example of a control system using negative feedback.y g g
• Phase-locked loops are widely used in radio, telecommunications, computers and other electronic applications. They may generate stable frequencies recover a signal from a noisy communication channel orfrequencies, recover a signal from a noisy communication channel, or distribute clock timing pulses in digital logic designs such as microprocessors. Since a single integrated circuit can provide a complete phase locked loop building block the technique is widely used in modernphase-locked-loop building block, the technique is widely used in modern electronic devices, with output frequencies from a fraction of a cycle per second up to many gigahertz.
Phase-locked loops (chapter 16) Low-pass
filt
Phasedetector
Outputvoltage
VinVpd
Klp
Gainfilter
Hlp s( )Vlp
Vcntl
Average voltage proportional to phase difference
• Clock multiplication:
VCOVcntl
Vosc(voltage controlled oscillator)
Clock multiplication:• The input signal is reference oscillator with fixed frequency• The PLL output is a signal with frequency N times the input frequency where N is an integer
• Data recovery and clock resynchronization:• The input signal is a digital signal containing data• The output is digital data at a certain clock rate• The system clock is recovered from the digital input signal
• Frequency synthesis (ex: to select channels in television or wireless communication systems):• The input signal is reference oscillator with fixed frequency• The PLL output is a signal with frequency N times the input frequency where N may be a fractional• The PLL output is a signal with frequency N times the input frequency where N may be a fractional
number• FM demodulation:
• The input is a FM signal (IF)• The output is the demodulated baseband signal
Phase-Locked Loop, typical architecture
• The phase detector (”PD”) normally has an output voltage with an average value proportional to the phase difference between the input signal and the output of the VCO (”Voltage Controlled Oscillator”).
• The low-pass filter is used to extract the average value from the output ofThe low pass filter is used to extract the average value from the output of the PD.
• The average value is amplified by the Gain block and used to drive the VCOVCO.
• The negative feedback of the loop results in the output of the VCO being synchronized with the input signal. 23. mars 2010 24
Phase-locked loop example (p. 649)
23. mars 2010 25
Vin and Vosc exactly in phase when Фd=90º
• Vin = Ein sin(ωt), • Vosc = E sin(ωt – Ф + 90º) = E sin(ωt -90º+ 90º) = E sin(ωt)• Vosc = Eosc sin(ωt – Фd + 90 ) = Eosc sin(ωt -90 + 90 ) = Eosc sin(ωt)• The input signal and the output of the oscillator will be exactly in phase, and the
output from the phase detector (” PD”) is found to be the amplitudes of the two i id l i li d h d di id d b 2 l i isinusoids multiplied together and divided by 2, resulting in
Vcntl = KlpKMEinEosc/2 (Eq. 16.5, p 650 in J&M)Fig. 16.2, below, shows the output from the PD when Vin and Vosc are nearly in phase.
23. mars 2010 26
Examples of different waveforms, as a function of different Фd,(J&M p 650-651 ,fig. 16.3)( p , g )
• Vosc = Eoscsin(ωt – Фd + 90°) = Eosccos(ωt – Фd) (16.4)• Фd > 0 corresponds to waveforms that are more in phase.Фd 0 corresponds to waveforms that are more in phase.• When the input signal and VCO output have a 90º phase diff. (Фd = 0),
they are uncorrelated and the output of the LP-filter will be zero.)
23. mars 2010 27
Some relevant mathematical relationships from 16.1, pages 650-652 in ”J&M”
• The output of the phase detector:• Vpd = KMVinVosc = KMEinEosc sin(ωt)cos(ωt-Фd) (16.6)• Using sin(A)cos(B) = (1/2)[sin(A+B)+sin(A-B)] we haveUsing sin(A)cos(B) (1/2)[sin(A+B)+sin(A-B)] we have
Vpd = KMVinVosc/2 [sin(Фd)+2sin(2ωt-Фd)] (16.8)• The LP-filter removes the second term at twice the frequency of the input signal,
V i th f i bso Vcntl is therefore given by Vcntl = KlpKM(EinEosc/2) sin(Фd) (16.9)
• Since Vcntl is either a dc value or slowly varying, for small Фd the following approximation for (16.9) may be used:Vcntl ≈Klp KM(EinEosc/2) Фd = KlpKpd Фd (16.10)Thus the output of the LP filter is approximately proportional to the phaseThus, the output of the LP-filter is approximately proportional to the phase difference between the output of the oscillator and the input signal, assuming the 90 ofset bias is ignored. The approximation is used in analyzing the PLL to obtain
li d l Th t t f ti lit i ll d K d i i ba linear model. The constant of proportionality is called Kpd and is given byKpd = KM(EinEosc/2) (16.11)
23. mars 2010 28
More on PLL operation (p. 652-653)• Assume that the VCO has a free-running
frequency ωfr when it’s input is zero, and that the input signal is initially equal to ωfr, and the system has Фd = 0 (in lock). y d ( )
• NEXT, assume the input frequency slowly increases. Now, with Фd > 0, the two waveforms will become more in phase (Seewaveforms will become more in phase (See fig. 16.3). After a short time the output of the LP-filter will go positive. Since the two
f t li htl diff twaveforms are at slightly different frequencies, the output of the LP-filter will slowly increase. Since the VCO frequency is proportional to Vcntl, this increase will cause the VCO frequency to increase until it is the same of that of the input signal again, which p g gwill keep the two signals in synchronism (i.e. locked). (The opposite would occur for a decrease in the input signal frequency.)23. mars 2010 29p g q y )
More on PLL operation (p. 652-653)• At a new input frequency, (≠ ωfr ) which
does not equal the free-running frequency, we can find the new phase difference for the two locked signals by noting that the frequency of g y g q ythe oscillator’s output signal is given by ωosc = KoscVcntl + ωfr (16.12).K is a constant relating the change inKosc is a constant relating the change in frequency to control voltage ratio. The output voltage of the amplified LP-filter is now
i bgiven by Vcntl = (ωinωfr)/Kosc (16.13)where ωin is the frequency of the input signal, which is equal to the frequency of the oscillator’s output. From (16.10): Фd = Vcntl(KlpKpd) = (ωin-ωfr)/KlpKpdKoscd cntl( lp pd) ( in fr) lp pd osc
(16.14)23. mars 2010 30
Ex. 16.2
23. mars 2010 31
Ex. 16.1
23. mars 2010 32
Linear model of the PLL – when in locklock
in s KlpHlp s
KpdVcnt l
osc s 1 s
Koscosc
Vcntl s Kpd KlpHlp s in s osc s – =
K V tl s osc s
KoscVcntl s s
------------------------------=
PLL transfer functions• Combining equations from previous slide:
V K K H ( )Vcntl s in s
-------------------sKpdKlpHlp s( )
s KpdKlpKoscHlp s( )+-----------------------------------------------------=
• This is a highpass response from input phase to the control voltagethe control voltage.
• Rewriting gives: K K K H s( )osc s in s ----------------
KpdKlpKoscHlp s( )s KpdKlpKoscHlp s( )+----------------------------------------------------=
• This is a lowpass response from the input phase to output phase
Additional litterature• Walt Kester: Which ADC Architecture is right for your application?, Analog Dialogue, Analog
Devices, 2005.• Behzad Razavi: ”Design of Analog CMOS Integrated Circuits”, McGraw-Hill, reprint 2009.• Jimmy J. Cathey, Syed A. Nasar: Basic Electrical Engineering , Schaum’s Outlines, McGraw
Hill 1997.• Richard Schreier, Gabor C. Temes: Understanding Delta-Sigma Data Converters, IEEE Press /Richard Schreier, Gabor C. Temes: Understanding Delta Sigma Data Converters, IEEE Press /
Wiley Interscience, 2005• Tapio Saramaki et. Al: Multiplier-Free Decimator Algorithms for Super-Resolution Oversampled
Converters, IEEE International Symposium on Circuits and Systems, 1990.• Bernhard A Boser Bruce A Wooley: The Design of Sigma Delta Modulation Analog to Digital• Bernhard A. Boser, Bruce A. Wooley: The Design of Sigma-Delta Modulation Analog-to-Digital
Converters, IEEE Journal of Solid-State Circuits, December 1988.• Sudhir M. Mallya, Joseph H. Nevin: Design Procedures for a Fully Differential Folded-Cascode
CMOS Operational Amplifier , IEEE Journal of Solid-State Circuits, December 1989.• http://www.wikipedia.org (on PLLs)• Snorre Aunet: Second-Order Sigma Delta Modulator, Nordic VLSI, May 4, 1994.
Next Time, 13/4-10:,
• More from Chapter 16; PLLs• About report writing• About report writing
TRANSFER
23. mars 2010 37
23. mars 2010 38
top related